Endothelium‐dependent hyperpolarizations: Past beliefs and present facts

Abstract

Endothelium‐dependent relaxations are attributed to the release of various factors, such as nitric oxide, carbon monoxide, reactive oxygen species, adenosine, peptides and arachidonic acid metabolites derived from the cyclooxygenases, lipoxygenases, and cytochrome P450 monooxygenases pathways. The hyperpolarization of the smooth muscle cell can contribute to or be an integral part of the mechanisms underlying the relaxations elicited by virtually all these endothelial mediators. These endothelium‐derived factors can activate different families of K⁺ channels of the vascular smooth muscle. Other events associated with the hyperpolarization of both the endothelial and the vascular smooth muscle cells (endothelium‐derived hyperpolarizing factor (EDHF)‐mediated responses) contribute also to endothelium‐dependent relaxations. These responses involve an increase in the intracellular Ca²⁺ concentration of the endothelial cells followed by the opening of Ca²⁺‐activated K⁺ channels of small and intermediate conductance and the subsequent hyperpolarization of these cells. Then, the endothelium‐dependent hyperpolarization of the underlying smooth muscle cells can be evoked by direct electrical coupling through myoendothelial junctions and/or the accumulation of K⁺ ions in the intercellular space between the two cell types. These various mechanisms are not necessarily mutually exclusive and, depending on the vascular bed and the experimental conditions, can occur simultaneously or sequentially, or also may act synergistically.

• Cardiovascular diseases
• cell membrane potential
• EDHF
• endothelial dysfunction
• endothelium
• myoendothelial gap junctions
• potassium channels

Abbreviations
11,12,15‐THETA	=	11,12,15‐trihydroxyeicosatrienoic acid
12‐S‐HETE	=	12‐(S)‐hydroxyeicosatetraenoic acid
20‐HETE	=	20‐hydroxyeicosatetraenoic acid
ANP	=	atrial natriuretic peptide
BKCa	=	large conductance calcium‐activated potassium channel
BNP	=	brain natriuretic peptide
CaV	=	voltage‐dependent calcium channel
CGRP	=	calcitonin gene‐related peptide
GIRK	=	G protein‐regulated inward‐rectifier potassium channel
CNP	=	C‐type natriuretic peptide
Cx	=	connexin
EDHF	=	endothelium‐derived hyperpolarizing factor
EET	=	epoxyeicosatrienoic acid
HO	=	heme‐oxygenase
IKCa	=	intermediate conductance calcium‐activated potassium channel
IP	=	receptor prostacyclin receptor
K2P	=	two‐pore domain potassium channel
KATP	=	ATP‐sensitive potassium channel
KCa	=	calcium‐activated potassium channel
KIR	=	inwardly rectifying potassium channel
KV	=	voltage‐dependent potassium channel
L‐NAME	=	L‐nitro‐arginine‐methyl‐ester
NPR‐C	=	natriuretic peptide receptor C subtype
PKA	=	protein kinase A
PKG	=	protein kinase G
SHR	=	spontaneously hypertensive rat
SKCa	=	small conductance calcium‐activated potassium channel
SERCA	=	sarcoplasmic reticulum Ca2+ ATPase
STOC	=	spontaneous transient outward current
TRPV4	=	vanilloid transient receptor potential channel
VIP	=	vasoactive intestinal peptide

Introduction

The endothelial cells, being at the interface of the blood and the vascular wall, are involved in many different physiological functions, including metabolism, angiogenesis, hemostasis, inflammation, synthesis and degradation of the extracellular matrix, as well as the regulation of vascular permeability and tone. The endothelium maintains the balance between inhibition and promotion of the proliferation and migration of smooth muscle cells, between prevention and stimulation of the adhesion and aggregation of the platelets, between thrombogenesis and fibrinolysis, as well as between vasodilatation and vasoconstriction. Upsetting this tightly regulated balance leads to endothelial dysfunction 1,2.

Endothelial cells control the tone of the underlying vascular smooth muscle cells by releasing various relaxing and contracting factors 3. Furchgott and Zawadzki 4 were the first to show that the presence of the endothelial lining is required to observe the relaxation of an isolated artery in response to acetylcholine. The term ‘endothelium‐dependent relaxations’ was then coined although other endothelium‐derived vasoactive substances had been identified previously (e.g. prostacyclin (PGI₂), adenosine). Nevertheless, it was only after this seminal work that the essential role of the endothelial cells in controlling the tone of the underlying vascular smooth muscle was fully understood, and usually attributed to the production of nitric oxide (NO) by the endothelial NO synthase (NOS‐3) 1,4. However, in numerous blood vessels from different species, including the human, endothelium‐dependent relaxations are still observed in the presence of inhibitors of cyclooxygenases and NO synthases. These relaxations are generally associated with the hyperpolarization of the vascular smooth muscle cells and were first attributed to a noncharacterized endothelial factor termed EDHF for endothelium‐derived hyperpolarizing factor (Figure 1). The acronym ‘EDHF’ turned out to be confusing because it implies that a single diffusible substance, different from nitric oxide (NO) and PGI₂, mediates this type of endothelium‐dependent relaxation. In fact, numerous endothelium‐derived factors, including NO and PGI₂ themselves, but also reactive oxygen species, carbon monoxide, adenosine, peptides, and various metabolites of arachidonic acid can hyperpolarize the underlying smooth muscle. Additionally, another pathway, which does not involve the release of a factor but is associated with the hyperpolarization of both the endothelial and the vascular smooth muscle cells, contributes also to certain endothelium‐dependent relaxations 1,5.

Figure 1. Endothelium‐dependent relaxation and endothelium‐dependent hyperpolarization. A: Left: The isometric tension developed by an arterial ring is studied in an organ chamber filled with thermostated and oxygenated physiological salt solution. The inset shows recording traces of an experiment reproducing the seminal observation first made by Furchgott and Zawadzki 4. An isolated ring of canine coronary artery with endothelium is contracted with prostaglandin F_2α (PGF_2α). The addition of acetylcholine produces a concentration‐dependent and endothelium‐dependent relaxation. Right: Bradykinin produces a concentration‐dependent and endothelium‐dependent relaxation in canine coronary arterial rings. The cyclooxygenase inhibitor, indomethacin, does not affect the relaxation to bradykinin while the NO synthase inhibitor L‐nitro‐arginine (L‐NA) produces a significant shift to the right of the concentration‐response curves, indicating a contribution of NO in the overall relaxation mechanism. The combination of the two inhibitors produces a further shift, indicating that prostacyclin contributes to the mechanism of bradykinin‐induced endothelium‐dependent relaxation when the NO synthase is inhibited. However, even in the presence of the two inhibitors, bradykinin still evokes the complete relaxation of the isolated arteries (modified from 210). SNP = sodium nitroprusside. B: Left: Changes in tension and in membrane potential are recorded in an isolated canine coronary artery with an intracellular microelectrode. Right: Acetylcholine produces a relaxation and a hyperpolarization of the vascular smooth muscle only when the endothelial cells are present (modified from 211).

The hyperpolarization of the smooth muscle cells is a powerful way to produce relaxation. It decreases not only Ca²⁺ influx by reducing the open probability of voltage‐dependent Ca²⁺ channel (Ca_V) but also the release of Ca²⁺ from intracellular stores by decreasing the turnover of intracellular phosphatidylinositol and the Ca_V‐dependent activation of the sarcoplasmic reticulum 1,6,7.

Key messages

• The hyperpolarization of vascular smooth muscle cells is a powerful way to produce a relaxation, and numerous endothelium‐derived factors hyperpolarize the underlying smooth muscle cells.

• Another pathway, which does not involve the release of a factor (endothelium‐derived hyperpolarizing factor (EDHF)‐mediated response), is associated with the hyperpolarization of both the endothelial and the vascular smooth muscle cells and contributes also to endothelium‐dependent relaxations.

• The gene deletion of each key molecular component involved in these EDHF‐mediated responses is associated with alterations in arterial blood pressure.

• In humans, these responses occur in vivo and contribute to the regulation of basal blood flow and to the endothelium‐dependent relaxations in response to flow and certain neurohumoral mediators.

Arachidonic acid metabolites

Cyclooxygenases

PGI₂ is the principal metabolite of arachidonic acid produced by cyclooxygenase in the endothelial cells 8. This prostaglandin, a preferential agonist of prostacyclin (IP) receptors, is a potent antithrombotic and antiplatelet agent as well as a vasodilator substance. Although several types of G proteins are likely to be coupled to the IP receptor, in vascular smooth muscle cells the Gs‐adenylyl‐cyclase‐cyclic adenosine monophosphate (cAMP)‐PKA pathway is a preponderant signaling system responsible for vascular relaxation 9. The relaxation to PGI₂, or its synthetic analogues, is often associated with the concomitant hyperpolarization of the smooth muscle cells and can involve the opening of ATP‐sensitive (K_ATP) 10,11, voltage‐dependent (K_V) 12, inwardly rectifying (K_IR) 13, large conductance Ca²⁺‐activated K⁺ channels (BK_Ca) 14,15 and/or two‐pore domain (K_2P) K⁺ channel family 16. The activation of K⁺ channels following IP receptor stimulation relies on cAMP‐dependent and ‐independent pathways. The latter pathway could be associated with the direct activation of the K⁺ channel by Gs protein 17.

Therefore, in numerous vascular beds, PGI₂ can act as an endothelium‐derived hyperpolarizing substance 1,11 (Figure 2). However, inhibitors of cyclooxygenases abolish the basal and stimulated generation of PGI₂, indicating that its release cannot account for the responses observed in the presence of the combination of inhibitors of NO synthase and cyclooxygenase.

Figure 2. Endothelium‐dependent hyperpolarizations in the guinea‐pig carotid artery. A: Acetylcholine (ACh) induces an endothelium‐dependent hyperpolarization. B: In the presence of charybdotoxin (ChTX) plus apamin plus indomethacin plus L‐nitro‐arginine (L‐NA), acetylcholine no longer produces an endothelium‐dependent hyperpolarization but a small depolarization. Depol. = depolarization. C: In the presence of charybdotoxin (ChTX) plus apamin, acetylcholine produces an endothelium‐dependent hyperpolarization attributable to the release of NO and PGI₂. PGI₂ = prostacyclin. D: In the presence of charybdotoxin (ChTX) plus apamin plus indomethacin, acetylcholine produces an endothelium‐dependent hyperpolarization attributable to the release of NO. E: In the presence of charybdotoxin (ChTX) plus apamin plus L‐nitro‐arginine, acetylcholine produces an endothelium‐dependent hyperpolarization attributable to the release of PGI₂. F: In the presence of indomethacin plus L‐nitro‐arginine (L‐NA), acetylcholine produces an EDHF‐mediated response. This response is not affected by the subsequent addition of a second NO synthase inhibitor, L‐monomethyl‐arginine (L‐NMMA) plus a NO scavenger, hemoglobin (middle trace), or another NO scavenger carboxy‐PTIO (lower trace), indicating that residual and stored NO are not involved.

Cytochrome P450

Epoxyeicosatrienoic acids (EETs), derived from cytochrome P450 2C or 2J epoxygenases, play an important role in endothelium‐dependent hyperpolarizations and relaxations in various blood vessels. In many arteries, inhibitors of cytochrome P450 monooxygenases or antisense oligonucleotides directed against cytochrome P450‐2C8‐9 inhibit endothelium‐dependent vasodilator responses that persist in the presence of inhibitors of NO synthases and cyclooxygenases 18,19. EETs can be involved in non‐NO‐, non‐PGI₂‐mediated responses as a factor released by the endothelial cells or without being a diffusible factor per se, for instance by acting as essential endothelial intracellular second messengers 18. However, bradykinin, pulsatile stretch, and shear stress can release EETs, which diffuse toward and then hyperpolarize the vascular smooth muscle cells by activating BK_Ca19–24. The mechanism by which EETs activate BK_Ca in vascular smooth muscle cells is not completely understood. They do not directly activate these channels as shown for other fatty acids 25. EETs may interact with an intracellular ‘receptor’ to activate adenosine diphosphate (ADP)‐ribosyltransferase. This enzyme catalyzes the ADP ribosylation of stimulatory G‐protein G_sα promoting the phosphorylation and the activation of the Slo1 α subunit of BK_Ca26,27. EETs can also activate BK_Ca in much more indirect manners. Indeed, 11,12‐EET can increase the activity of heme‐oxygenase (HO) and the subsequent generation of carbon monoxide (CO) activates BK_Ca28. Furthermore, EETs can promote Ca²⁺ entry in smooth muscle cells by activating cationic channels such as the vanilloid transient receptor potential channel (TRPV4). The Ca²⁺ influx through TRPV4 preferentially stimulates ryanodine receptors located on the sarcoplasmic reticulum increasing the frequency of unitary Ca²⁺ release events (Ca²⁺ sparks). The EETs‐induced Ca²⁺ sparks activate nearby sarcolemmal BK_Ca and increase the frequency of spontaneous transient outward currents (STOCs). This EETs‐dependent activation of the Ca²⁺‐signaling complex (TRPV4‐ryanodine receptors‐BK_Ca) hyperpolarizes and relaxes vascular smooth muscle cells 29.

Therefore, in some vascular beds, especially but not exclusively in coronary arteries, and following appropriate stimulation, the endothelial cells can release EETs, which act as diffusible endothelium‐derived hyperpolarizing substances and contribute to relaxations resistant to inhibitors of NO synthases and cyclooxygenases. Additionally, cytochrome P450 metabolites play an important role in the regulation of kidney the circulation and contribute to the long‐term regulation of blood pressure and sodium homeostasis. EETs not only regulate regional blood flow, but possess anti‐inflammatory actions, improve insulin sensitivity and lipid metabolism that could protect the kidney vasculature from injury during renal and cardiovascular diseases and prevent/delay metabolic syndrome and atherogenesis 30,31.

However, the effects of EETs are specific for certain vascular beds and in many blood vessels EETs evoke either minor relaxations (or hyperpolarization) and do not contribute to endothelium‐dependent relaxations 1,32,33.

Lipoxygenase

The lipoxygenase metabolites 12‐(S)‐hydroxyeicosatetraenoic acid (12‐S‐HETE) and 11,12,15‐trihydroxyeicosatrienoic acid (11,12,15‐THETA) can be released by the endothelial cells and evoke relaxation of the vascular smooth muscle by activating K_Ca34,36. However, the involvement of lipoxygenase derivatives in endothelium‐dependent relaxations appears restricted to a limited number of blood vessels 1,37.

Endocannabinoids

In isolated blood vessels or in anaesthetized animals, endogeneous and exogeneous cannabinoids usually have vasodilator properties and are likely to play a role in the regulation of the cardiovascular system 38. However, the suggestion that anandamide could be an endothelium‐derived vasoactive factors released by neurohumoral mediators has not been substantiated 1,39.

Gaseous mediators

Gaseous mediators are not involved like other neurohumoral mediators in classic ligand‐membrane‐bound receptor interactions. They readily cross lipid membranes and therefore diffuse homogeneously, in a nonpolarized manner from their production site. Nitric oxide can be considered as the first member of this new family of neurohumoral substances. Two other low‐molecular‐weight compounds can be included in this family: carbon monoxide and hydrogen sulfide 40.

Nitric oxide (NO)

NO is a potent vasodilator and a powerful inhibitor of platelet adhesion and aggregation. The relaxations elicited by NO are generally associated with the stimulation of the cytosolic soluble guanylyl cyclase and the subsequent activation of cyclic guanosine monophosphate (cGMP)‐dependent protein kinase (PKG). However, the mechanisms underlying the relaxations caused by NO also involve cyclic GMP‐independent mechanisms such as the stimulation of the uptake of cytosolic Ca²⁺ via the sarcoplasmic reticulum Ca²⁺ ATPase (SERCA). The combined decrease in the Ca²⁺ sensitivity of the contractile proteins and the lowering of intracellular free Ca²⁺ explain the powerful vasodilatory properties of NO 41,42. NO regulates the activity of various ion channels and, depending on the vascular beds, the hyperpolarizing effects of NO on vascular smooth muscle cells can substantially contribute to their relaxation. NO can activate BK_Ca43,44 via a PKG‐dependent phosphorylation of the channel 45 or by a direct, cyclic‐GMP‐independent manner 46,47. NO can also activate K_V12,48, K_ATP49,50 and/or K_IR51. Beyond K⁺ conductances, NO also interacts with other ionic channels, such as chloride and cationic channels, and can therefore further influence the membrane potential of the smooth muscle cells 1.

In contrast to inhibitors of cyclooxygenase, inhibitors of NO synthase do not necessarily fully inhibit the production of NO. Thus, in their presence, residual NO can be produced by the endothelial cells and contributes to the relaxation and/or hyperpolarization of the underlying vascular smooth muscle 52,53. Furthermore, NO can also be stored and released independently of the activation of NO synthase. Two such stores have been described: the photosensitive store 54,55 and the adventitial dinitrosyl nonheme iron complexes and S‐nitrosated proteins 56,57.

Therefore, it is obvious that the role of NO as an endothelium‐derived hyperpolarizing substance has certainly been underestimated assuming that the presence of an inhibitor of NO synthases rules out its contribution. A non‐NO‐, non‐PGI₂‐mediated response should be reported not only when a relaxation and/or hyperpolarization is recorded in the combined presence of inhibitors of cyclooxygenases and NO synthases, but when this response is still observed with the additional presence of a NO scavenger 58 (Figure 2).

Carbon monoxide (CO)

The predominant biological source of CO is from the heme degradation by heme‐oxygenase (HO), which generates biliverdin, free ferrous iron, and carbon monoxide. Three isoforms of HO have been described HO‐1 (inducible isoform), HO‐2 (constitutive isoform), and HO‐3, nearly devoid of catalytic activity 59. Both HO‐1 and HO‐2 are expressed in vascular smooth muscle and endothelial cells 40,60. In many physiopathological situations the HO‐CO pathway compensates for the decrease NO bioavailability 59.

CO is a potent vasodilator in most, but not all, vascular beds. The mechanisms of CO‐induced vasodilatation involve the stimulation of soluble guanylate cyclase, the inhibition of cytochrome P450, and/or the activation of K⁺ channels 54,59. CO and CO‐donors activate smooth muscle BK_Ca, K_ATP61, and/or K_V 62. CO induces the cGMP‐dependent activation of BK_Ca but also directly increases the open probability of this potassium channel 1,59. CO augments the apparent Ca²⁺ sensitivity of the α subunit of BK_Ca, enhancing the probability of eliciting STOCs in response to Ca²⁺ sparks 63. Additionally, CO, like NO, can also indirectly activate BK_Ca by inhibiting the cytochrome P450‐dependent synthesis of 20‐HETE, the endogenous vasoconstrictor and tonic inhibitor of BK_Ca64. However, CO is also a tonic inhibitor of NOS by binding to its prosthetic heme and can contribute to endothelial dysfunction 65.

An endothelial production of CO, contributing to endothelium‐dependent relaxations in response to neurohumoral substances, has been demonstrated only in a limited number of arteries 1,66,67.

Hydrogen sulfide (H₂S)

Two main enzymes are responsible for the production of hydrogen sulfide (H₂S), cystathionine β‐synthase and cystathionine γ‐lyase and both use L‐cysteine as substrate. The physiological cardiovascular effects of H₂S are linked to the latter enzyme, which in the vascular wall is expressed mostly in smooth muscle but is virtually absent in endothelial cells 40. Therefore, H₂S does not qualify as an endothelium‐derived mediator.

Reactive oxygen species

Hydrogen peroxide (H₂O₂)

H₂O₂, depending on the blood vessel, the presence of the endothelium, the experimental conditions, or the concentrations studied, possesses dilator or constrictor properties. The endothelial production of hydrogen peroxide (H₂O₂) can be involved in endothelium‐dependent relaxations in response to agonists and flow 68,69 or compensate for the decreased production of NO 70. In some arteries, endothelium‐derived H₂O₂ evokes the hyperpolarization and relaxation of smooth muscle cells 71–73, and in vivo this reactive oxygen species could play an important role in coronary autoregulation 74 and a cardioprotective role during ischemia‐reperfusion injury 75.

The endothelium‐independent relaxations of arterial myocytes to H₂O₂ involve multiple pathways, direct effects such as the activation of guanylyl cyclase and/or K⁺ channels, and indirect effects via the metabolism of arachidonic acid either through cyclooxygenases or lipoxygenases 76. H₂O₂ produces the hyperpolarization of the vascular smooth muscle cells by activating BK_Ca77,78, K_ATP79, K_V80, and/or K_IR channels 81. The activation of smooth muscle K_V by H₂O₂ involves redox‐sensitive elements, thiol groups, on the channel itself or on associated regulatory proteins 80. The activation of BK_Ca by H₂O₂ is evoked by a direct action on the channel as well as by a soluble guanylyl cyclase‐dependent effect 77. However, H₂O₂ can also be a potent inhibitor of BK_Ca by altering two distinct cysteines situated on its α subunit 82.

Taken in conjunction, the available evidence suggests that H₂O₂ has multiple effects on the vascular cells. Nevertheless, as this reactive oxygen species is produced by the endothelial cells and in some arteries evokes hyperpolarization and relaxation of smooth muscle cells, H₂O₂ can be an endothelium‐derived hyperpolarizing mediator 73. However, H₂O₂ does not always hyperpolarize vascular smooth muscle cells 83, and its relaxing effect is not necessarily associated with hyperpolarization of the vascular smooth muscle cells or/and the opening of K⁺ channels 84. Furthermore, in many arteries EDHF‐mediated responses cannot be attributed to the generation of H₂O₂1,76.

Other reactive oxygen species

Hydroxyl radicals are endothelium‐derived relaxing factors in mouse cerebral arterioles 85. In vascular smooth muscle cells, superoxide anions may activate K_Ca in a cGMP‐independent manner 79, whereas peroxynitrite can in some arteries activate K_ATP and/or K_Ca, possibly by a cGMP‐dependent mechanism 86.

Vasoactive peptides

Endothelial cells can theoretically synthesize numerous vasoactive peptides. Among them C‐type natriuretic peptide (CNP), vasoactive intestinal peptide (VIP), calcitonin gene‐related peptide (CGRP), and adrenomedullin cause relaxation of vascular smooth muscle cells.

CNP

The natriuretic peptide family includes four distinct peptides, atrial natriuretic peptide (ANP), brain natriuretic peptide (BNP), CNP and urodilatin which are involved in the control of body fluid homeostasis and blood pressure. In contrast to ANP and BNP that are synthesized predominantly in the heart, CNP is produced in the central nervous system but has also been detected in peripheral endothelial cells and to a lesser extent in vascular smooth muscle cells 87,88. CNP evokes relaxations and hyperpolarizations of arterial and venous smooth muscle cells 89,90. This natriuretic peptide induces the accumulation of cGMP and opens BK_Ca91–93.

Additionally, in the rat mesenteric artery, endothelium‐derived CNP could activate natriuretic peptide receptor (NPR)‐C subtype present in smooth muscle. Hyperpolarization of the smooth muscle cell would be evoked via the cyclic‐GMP‐independent activation of a G protein‐regulated inward‐rectifier K⁺ channel (GIRK) 94,95. This latter hypothesis, based on pharmacological evidence, suggesting that CNP is an endothelium‐derived hyperpolarizing substance, should be examined with caution and awaits confirmation with experiments associating electrophysiological and molecular biology techniques 53,96,97. In the isolated guinea‐pig carotid artery, endothelium‐dependent hyperpolarizations do not rely on the release of CNP 98.

Other peptides

Adrenomedullin, CGRP, VIP and pituitary adenylyl cyclase‐activating polypeptide are extremely potent vasodilator peptides, and the mechanism underlying the relaxation often involves the hyperpolarization of the vascular smooth muscle cells. However, there is very little evidence so far that these peptides can be released acutely by endothelial cells. Since the three latter are well established neurotransmitters, they could be considered as ‘nerve‐derived hyperpolarizing factors’ 1.

Adenosine

Adenosine, an endogenous vasodilator released from contracting skeletal muscle and myocardium, contributes to exercise hyperemia and is an important regulator of coronary blood flow especially during ischemia 99,100. Endothelial cells can also release adenine nucleotides and adenosine, but their contribution appears modest when compared to that of cardiac or skeletal muscle cells. Endothelial cells, in response to various stimuli including hypoxia and increases in shear stress, release mainly adenosine triphosphate (ATP), but this nucleotide is rapidly transformed into adenosine by ectonucleotidases. However, in most blood vessels there is little evidence that adenosine plays a major role as an endothelium‐derived relaxing or hyperpolarizing factor 1.

Endothelium‐dependent hyperpolarizations (EDHF‐mediated responses)

EDHF‐mediated responses are endothelium‐dependent relaxations resistant to inhibitors of NO synthases and cyclooxygenases, which do not involve one of the mediators mentioned above (residual or stored NO, EETs, CNP, H₂O₂, and so on). These relaxations are produced by agonists that stimulate G protein‐coupled receptors, by substances that increase the endothelial intracellular Ca²⁺ concentration and by physical stimuli such as shear stress 1,58.

Involvement of potassium channels

EDHF‐mediated responses are insensitive to specific blockers of K_ATP and BK_Ca. In most blood vessels, they involve the activation of small conductance K_Ca (SK_Ca), blocked by apamin, scyllatoxin, or UCL 1684, and/or intermediate conductance K_Ca (IK_Ca), blocked by charybdotoxin or TRAM‐34 33,50,101,102 (Figure 3). In freshly isolated vascular smooth muscle cells from various species, IK_Ca and SK_Ca are not or very poorly expressed whereas functional BK_Ca are constitutively present 103,104. In contrast, both SK_Ca (and especially the SK3 α subunit) and IK_Ca (IK1) are expressed in freshly isolated endothelial cells from various arteries. These channels are activated by substances that elicit EDHF‐mediated responses, and their activation produces the hyperpolarization of the endothelial cells 105–109. Thus, the two K⁺ channels involved in EDHF‐mediated responses, SK_Ca and IK_Ca, are located on the plasma membrane of the endothelial cells.

In anesthetized rats, the acute administration of charybdotoxin plus apamin does not affect basal arterial blood pressure but attenuates the decrease in blood pressure produced by acetylcholine or ghrelin 110,111. Similarly, in the mesenteric and hindlimb vascular beds, the combination of the two toxins does not affect the basal conductance but inhibits the vasodilatation evoked by acetylcholine or bradykinin, indicating that EDHF‐mediated responses play a physiological role in the regulation of vascular resistance 112.

Beyond endothelial hyperpolarization

In the vascular wall, gap junctions link smooth muscle with other smooth muscle cells, endothelial with other endothelial cells, and in many blood vessels smooth muscle with endothelial cells. The number of myoendothelial gap junctions increases with a reduction in the size of the artery 113, a phenomenon that parallels the contribution of the EDHF‐mediated responses to endothelium‐dependent relaxations 114,115. Furthermore, the presence of these heterocellular gap junctions is associated with EDHF‐mediated responses 116,117. Endothelium and smooth muscle cells can communicate via these myoendothelial gap junctions physically, as Ca²⁺ can diffuse from one cell type to another 118,119, and electrically, since depolarization and hyperpolarization are conducted bidirectionally from one cell type to the other 120–126. The connexins (Cx) 37, 40, and 43 are the predominant isoforms of gap junction proteins expressed in the vascular wall. Depending on the blood vessel, they can be involved in myoendothelial gap junction communication 127–130. Blockers of gap junctions abolish or partially inhibit EDHF‐like responses in many arteries in vitro127,131–136 and in vivo137. Furthermore, in the rat mesenteric artery, antibodies directed against Cx40, when loaded selectively in the endothelial cells block EDHF‐mediated responses 138. In this preparation, SK_Ca colocalize with Cx37, Cx40, and Cx43 and are adjacent to endothelial cell gap junctions, whereas IK_Ca colocalize with Cx37 and Cx40 in myoendothelial gap junctions 139. Therefore, the transmission of endothelial hyperpolarization to the underlying smooth muscle via myoendothelial gap junctions is likely to play a predominant role in EDHF‐mediated responses in many blood vessels. Additionally, the endothelial cell lining is a privileged low‐resistance pathway for conducted vasodilatation/hyperpolarization along the length of the vessel 124,125.

An additional mechanism to achieve hyperpolarization and relaxation of the underlying vascular smooth muscle cells is also directly linked to the hyperpolarization of the endothelial cells. Indeed, the activation of endothelial IK_Ca and SK_Ca causes an efflux of K⁺ ions from the intracellular compartment toward the extracellular space. K⁺ ions accumulate in the intercellular space between endothelial and smooth muscle cells. A moderate increase in the extracellular K⁺ concentration (1 to 15 mM) can provoke the relaxation of vascular smooth muscle cells 140 by activating K_IR141 and the Na⁺/K⁺ pump 142. Indeed, a unique feature of K_IR channels is the action of extracellular potassium on its gating. A moderate increase in potassium concentration, in the range of 1 to 20 mM, enhances potassium efflux through K_IR at physiologically relevant potentials 143. The activation of K_IR and the Na⁺/K⁺ pump overcomes the small depolarizing effects linked to the increase in potassium ions per se and the net resultant is hyperpolarization and thus relaxation of the smooth muscle cells. The K_IR channel can be regarded as a metabolic sensor producing vasodilatation and increases in blood flow when potassium accumulates in the circulation during neuronal activity or exercise for instance. Therefore, accumulation of K⁺ ions in the intercellular space may contribute to the mechanism of EDHF‐mediated responses. This hypothesis was demonstrated successfully in the hepatic and mesenteric arteries of the rat 144. Edwards et al. provided evidence that the targets of apamin and charybdotoxin are located on the endothelial cells, that potassium ions accumulate in the intercellular space between endothelial and smooth muscle cells, and that the potassium efflux associated with the activation of endothelial IK_Ca and SK_Ca channels produced hyperpolarization by activating both K_IR and the Na⁺/K⁺ pump on the smooth muscle cells. A role for the accumulation of K⁺ ions in the intercellular space was then confirmed in many different arteries 145–152. Nevertheless, this hypothesis has been highly controversial, possibly because the experimental conditions, and especially the presence or not of contractile agents, can markedly influence the contribution of K⁺ ions in the overall mechanisms underlying endothelial‐dependent hyperpolarizations. Contractile agonists stimulate receptors on vascular smooth muscle cells, increase intracellular calcium concentration, and depolarize the cells. The rise in intracellular calcium activates BK_Ca while the depolarization activates K_V as braking mechanisms. Potassium effluxes through these two channels, and K⁺ ions accumulate in the intercellular space. This ‘potassium cloud’ activates K_IR and Na⁺/K⁺‐ATPase, preventing the effects of any subsequent rise of potassium during endothelium‐dependent hyperpolarization 153.

The K_IR channel, most likely involved in K⁺ ion‐induced hyperpolarization in rat arteries, is composed of the Kir2.1 α subunits 154–156. However, in some blood vessels, K⁺ does not evoke or inconsistently produces relaxations and hyperpolarizations 126,134,135,157,158. Therefore, in these blood vessels the contribution of K⁺ ions in EDHF‐mediated responses must be, if anything, minimal.

The involvements of gap junctions and K⁺ ions are not necessarily mutually exclusive. The relative proportion of each mechanism almost certainly depends on numerous parameters including the state of activation of the vascular smooth muscle, the density of myoendothelial gap junctions, and the level of the expression of the appropriate isoforms of Na⁺/K⁺‐ATPase and/or K_IR on the vascular smooth muscle 1,58 (Figure 3).

Figure 3. EDHF‐mediated responses. The activation of endothelial receptors (R) and the shear stress exerted by the flowing blood increase endothelial [Ca²⁺]_i through calcium entry (TRP = transient receptor potential channel) and calcium release from internal stores (SR = sarcoplasmic reticulum). This activates endothelial K_Ca channels and produces endothelial hyperpolarization which is conducted through myoendothelial gap junctions to the underlying vascular smooth muscle. Cyclic‐AMP enhances the electrotonic conduction via gap junctions. Additionally, accumulation of potassium ions in the intercellular space can hyperpolarize the smooth muscle cells by activating K_IR and Na⁺/K⁺‐ATPase. The presence of a ‘potassium cloud’ can markedly affect the contribution of K⁺ ions to EDHF‐mediated responses. Contractile agonists that stimulate their receptor on vascular smooth muscle cells increase intracellular calcium (Ca²⁺) concentration and depolarize the cells. The rise in intracellular calcium activates BK_Ca while the depolarization activates K_V as a braking mechanism. Potassium effluxes through these two channels and accumulates in the intercellular space (potassium cloud). Potassium ions activate K_IR and Na⁺/K⁺‐ATPase, preventing the effects of any subsequent rise of potassium during endothelium‐dependent hyperpolarization. The relative proportion of each mechanism depends on numerous parameters, including the state of activation of the vascular smooth muscle, the density of myoendothelial gap junctions, and the level of the expression of the appropriate isoforms of Na⁺/K⁺‐ATPase and/or K_IR. BK = bradykinin; SP = substance P; SLIGRL‐NH2 = peptide agonist of proteinase‐activated recepor‐2; P3 = inositol trisphosphate; TRPV4 = transient receptor potential channel vanilloid 4; AC = adenylyl cyclase; αGA = glycyrrhetinic acid derivatives; CX = connexin; 4‐AP = 4‐aminopyridine; IbTX = iberiotoxin; ChTX = charybdotoxin; SK3 = small conductance calcium‐activated potassium channel formed by SK3 α subunits; IK1 = intermediate conductance calcium‐activated potassium channel formed by IK1 α subunits; Kir2.1 = Inward rectifying potassium channel constituted of Kir2.1 α subunits; K_V = voltage‐gated potassium channels; BK_Ca = large conductance calcium‐activated potassium channels (modified from 58); ACh = acetylcholine.

Genetically modified animals

In NOS‐3 knockout mice, EDHF‐mediated responses play a compensatory role for the absence of endothelial NO 159 (Figure 4), and this adaptation to NO synthase deletion is gender‐specific 160. Similarly, in resistance arteries from double‐knockout mice for NOS‐3 and cyclooxygenase‐1 (COX‐1), endothelium‐dependent relaxations are preserved by an EDHF‐mediated mechanism, while in arteries from the males the endothelium‐dependent relaxations are impaired severely. In those female mice, the double deletion of NOS‐3 and COX‐1 does not affect mean arterial blood pressure, while the corresponding males are hypertensive 161.

Figure 4. EDHF‐mediated responses in two strains of mice, wild‐type (C57BL/6), and NOS‐3 knockout mice (eNOS(−/−)). A: NOS‐3 knockout mice are hypertensive. B: Acetylcholine (ACh) produces an endothelium‐dependent relaxation in the isolated aortic rings of wild‐type mice (eNOS(+/+)) but no longer in that of NOS‐3 knockout mice (eNOS(−/−)). However, the NO donor and endothelium‐independent vasodilator sodium nitroprusside (SNP) produces a relaxation in the aorta from both strains. These results confirm the deletion of the NOS‐3 gene and demonstrate that, in a large blood vessel such as the aorta, NO is the preponderant endothelium‐derived relaxing factor released by acetylcholine. C: Acetylcholine induces a vasodilatation in the hindlimb of both wild‐type and NOS‐3 knockout mice. In the wild‐type mice, this vasodilatation is unaffected by a depolarizing solution (KCl) and partially inhibited by the combination of charybdotoxin (CTX) and apamin (apa) but not by charybdotoxin alone, indicating that an EDHF‐mediated response may contribute to the mechanism of vasodilatation but is not the preponderant mechanism. In the NOS‐3 knockout mice, the acetylcholine‐induced vasodilatation is preserved and is blocked by KCl or the combination of charybdotoxin plus apamin, partially inhibited by charybdotoxin alone and unaffected by the combination of iberiotoxin (IbTX), apamin, glibenclamide (Glib) and 4‐aminopyridine (4AP). These results indicate that acetylcholine induces an endothelium‐dependent vasodilatation showing the hallmarks of EDHF‐mediated responses. The EDHF‐mediated response compensates the disruption of the NOS‐3 gene. D: Original recording of arterial blood pressure in an anaesthetized NOS‐3 knockout mouse treated with an inhibitor of cyclooxygenase, diclofenac, and an inhibitor of NO synthase, L‐arginine‐methyl‐ester (L‐NAME). The intravenous administration of the endothelium‐dependent vasodilator bradykinin induces a transient decrease in arterial blood pressure. The summary bar graphs show the concentration‐dependent changes in mean arterial pressure (MAP) evoked by administration of bradykinin in wild‐type and eNOS(−/−) mice under control conditions (top graphs) or after treatment with inhibitors of cyclooxygenases and NO synthases (lower graphs). The amplitude of the hypotensive response induced by bradykinin is similar in NOS‐3 knockout and wild‐type animals. In both strains, the effect of bradykinin is unaffected by the presence of inhibitors of cyclooxygenases and NO synthases. These data strongly suggest that EDHF‐mediated responses play an essential role in the hypotensive effect of bradykinin (modified from 159). CTL = control.

Transgenic mice harboring genetically targeted alleles for the SK3 channel, in which SK3 gene expression can be experimentally controlled by dietary doxycycline, have been engineered 162. In those transgenic mice, the suppression of SK3 expression in the endothelial cells depolarizes both the endothelial and the vascular smooth muscle cells, reduces the diameter of resistance vessels in situ, and increases arterial blood pressure 108. Disruption of the IK1 gene reduces the hyperpolarization of endothelial and smooth muscle cells in response to acetylcholine. The associated vasodilatation in the carotid artery and in resistance vessels is also diminished because of a substantial reduction in EDHF‐mediated responses. Moreover, the IK1 deletion also led to a significant increase in arterial blood pressure and to mild left ventricular hypertrophy 109. These results confirm that in mice the endothelial SK_Ca and IK_Ca are fundamental determinants of endothelial hyperpolarization and EDHF signaling, and indicate that they actively control vascular tone and the overall regulation of the circulation. In mice that lack Kir2.1, but not in those deleted for Kir2.2 genes, K_IR currents are absent, and stimulation with moderate increase in K⁺ concentration does not produce relaxation, indicating that Kir2.1 gene expression is required for K_IR currents and K⁺‐induced dilatation. Kir2.1(−/−) animals die 8–12 hours after birth preventing the acquisition of more information on cardiovascular parameters such as arterial blood pressure 154.

Connexins 37 and 40 are the predominant gap junction proteins in the endothelial cells of the mouse 163. Connexin 40 proteins are involved in endothelial homocellular gap junctions and also in heterocellular gap junctions linking endothelial cells not only to smooth muscle cells but also to renin‐producing juxtaglomerular cells. The presence of the latter gap junction communication is required in order to maintain the calcium‐dependent inhibitor effects of angiotensin II and that of intrarenal pressure on renin secretion and synthesis, suggesting that the endothelium is strongly involved in the regulation of the renin system. Mice deficient for connexin 40 are hypertensive. However, the control of renin production only partially explained the hypertension in connexin 40 knockout mice 164. The arterioles of these animals also exhibit a reduced spread of dilatation in response to endothelium‐dependent vasodilators and irregular arteriolar vasomotion 165–168. Mice subjected to specific deletion of endothelial connexin 43 do not present major alterations in arterial blood pressure 169,170, possibly because this connexin is not the major one expressed in murine endothelial cells.

These results show that deletion of each key molecular component of EDHF‐mediated responses is associated with hemodynamic alterations suggesting that this endothelial pathway contributes to the overall regulation of arterial blood pressure.

Endothelium‐dependent responses resistant to inhibitors of NO‐synthases and cyclooxygenases in human blood vessels

Isolated blood vessels

In isolated human vascular preparations, endothelium‐dependent responses resistant to inhibitors of NO synthases and cyclooxygenases are observed in coronary (Figure 5), internal mammary, radial, renal, and cerebral arteries, as well as in resistance arteries from various vascular beds and in forearm veins 1. As in animal arteries, the contribution of EDHF‐mediated responses appears significantly more important in smaller than in large human arteries 171,172.

Figure 5. Endothelium‐dependent hyperpolarizations in isolated human coronary arteries (in the presence of inhibitors of cyclooxygenases and NO synthases). A: In coronary arteries taken from a young patient (15 months old, congenital heart disease), bradykinin produces an endothelium‐dependent hyperpolarization. B: In coronary arteries taken from an older patient (64 years old, ischemic heart disease) the endothelium‐dependent hyperpolarization to bradykinin is reduced markedly when compared to that observed in the juvenile patient. C: In coronary arteries taken from a patient (68 years old, ischemic cardiomyopathy) the endothelium‐dependent hyperpolarization to bradykinin is virtually absent. The presence of the angiotensin‐converting enzyme inhibitor, perindoprilat, restores the endothelium‐dependent hyperpolarization (modified from 1).

These hyperpolarizations or/and relaxations resistant to inhibitors of NO synthases and cyclooxygenases can be attributed to the activation of K⁺ channels as they are blocked by elevated K⁺ concentrations. These responses are independent of K_ATP activation and in general are minimally or not at all affected by charybdotoxin, iberiotoxin, or apamin given alone but are blocked either by nonspecific K⁺ channel inhibitors or by the combination of charybdotoxin plus apamin. The various hypotheses formulated for the identification of these responses are as follows: 1) EETs derived from cytochrome P450 monooxygenase in coronary, internal mammary, and subcutaneous arteries 173–176; 2) H₂O₂ in coronary and mesenteric arteries 68,72; and 3) EDHF‐mediated responses with a preferential involvement of Cx43‐dependent myoendothelial gap junction communication in subcutaneous arteries 126,128,177 and a preferential involvement of K⁺ ions in interlobar renal and thyroid arteries 152,178. CNP also relaxes and hyperpolarizes human subcutaneous resistance arteries by a cyclic GMP‐dependent activation of BK_Ca179, but its contribution to endothelium‐dependent responses has not yet been demonstrated. To the best of our knowledge, there is no conclusive evidence for a major role of lipoxygenase in the endothelium‐dependent relaxations/hyperpolarizations of human blood vessels. Contractile lipoxygenase derivatives, generated in response to various agonists, such as angiotensin II, arachidonic acid, acetylcholine, or bradykinin, can play a role in many blood vessels, but they do not originate from the endothelial cells 180.

Thus, the in vitro data available demonstrate that endothelium‐dependent responses resistant to inhibitors of NO‐synthases and cyclooxygenases are observed in isolated human blood vessels and that the characteristics of these responses are similar to those observed in most animal studies (Table I).

Table I. Nonexhaustive summary of potential mechanisms underlying endothelium‐dependent hyperpolarizations/relaxations resistant to the combined presence of inhibitors of NO‐synthases and cyclooxygenases.

Species	Artery	Stimulus	NO	MEGJ	K^+ ions	EETs	Lipox.	H2O2	CNP	References
Guinea‐pig	Carotid	ACh		✓						134
	Mesenteric	ACh		✓						122,123
		SP		✓						126
	Coronary	ACh		✓						212
Rat	Hepatic	ACh	✓	✓	✓					55,132,144
	Mesenteric	ACh	✓	✓	✓				✓	55,94,113,130,134,138,144
		Flow						✓		213
	Cerebral	ATP, UTP		✓						214
		SLIGRL			✓					215
	Skelet. M.	Flow				✓				23
		ACh		✓		✓				216
Mouse	Mesenteric	ACh		✓	✓			✓		71,117,217
		SLIGRL		✓	✓					147
	Coronary	BK						✓		218
	Hindlimb	ACh		✓						159
Rabbit	Aorta	ACh					✓			35
	Carotid	ACh	✓			✓				52,219,220
	Ear	ACh		✓						136
	Iliac	ACh		✓						221
	Mesenteric	ACh		✓	✓		✓			133,222
	Cerebral	ACh		✓						223
Pig	Coronary	BK	✓	✓	✓	✓		✓		19,24,53,57,148,224
		SP		✓	✓			✓		24,148,224,225
	Renal	BK			✓					151
Dog	Coronary	ACh				✓		✓		74,226,227
Cow	Coronary	Methachol.				✓				21
		BK		✓	✓	✓				22,149,228
Human	Coronary	Arachid. A.				✓				173
		Flow						✓		68
	Mammary	ACh				✓				176
		BK				✓				176
	Radial	Flow				✓				187
	Mesenteric	BK								72
	Renal	ACh			✓			✓		152
	Resistance	BK		✓						128,177,229
		ACh			✓	✓				126,175,178

This table must be read with a critical mind. Firstly, the authors of the various reported studies very often only suggest a potential mechanism in order to explain these endothelium‐dependent hyperpolarizations/relaxations that are resistant to the combined presence of inhibitors of NO‐synthases and cyclooxygenases. Secondly, specific tools were not necessarily available at the time when some of the above‐mentioned studies were performed, possibly leading to erroneous conclusions. Thirdly, for the sake of clarity, this table has been simplified, and for instance a subheading such as ‘coronary artery’ includes data involving large and small arteries that may exhibit very different behavior.

NO = residual and/or stored nitric oxide; MEGJ = myoendothelial gap junction communication; EETs = cytochrome P450‐derived epoxyeicosatrienoic acids; Lipox = lipoxygenase derivatives; Skelet. M. = skeletal muscle; CNP: C‐type natriuretic peptide; Ach = acetylcholine; BK = bradykinin; SP = substance P; SLIGRL = tethered ligand sequence for proteinase‐activated receptor‐2 (PAR‐2); Arachid. A. = arachidonic acid; Methachol. = methacholine; ATP = adenosine triphosphate; UTP = uridine triphosphate.

Intact organism

An EDHF‐mediated mechanism is often suggested to explain vasodilatations resistant to inhibitors of NO synthase. However, many human studies do not involve the administration of an inhibitor of cyclooxygenases. Furthermore, complete blockade of NO synthase is difficult to obtain (or to demonstrate), and/or a nonendothelial effect of the vasodilators such as a direct effect on the smooth muscle cells or an inhibitory effect on the sympathetic system cannot be excluded easily. In the human, apamin plus charybdotoxin cannot be administered for obvious ethical reasons. However, studies especially designed to study EDHF‐mediated responses have shown convincingly that in the forearm vasculature, basal forearm blood flow, flow‐induced vasodilatation, as well as the increases in blood flow in response to bradykinin, substance P, and to a lesser extent acetylcholine involve a non‐PGI₂‐, non‐NO‐mediated mechanism that depends on activation of K⁺ channels 181–188.

The mechanism underlying the non‐NO‐, non‐PGI₂‐mediated responses is still controversial. The endothelial production of a cytochrome P450 metabolite has been suggested to play a role in the regulation of the basal diameter and flow‐induced vasodilatation of the radial artery 187,188 as well as in exercise‐induced increases in skeletal muscle blood flow 189. Similarly, in hypertensive patients, a sulfaphenazole‐sensitive component contributes to the increase in forearm blood flow evoked by the administration of bradykinin and to a lesser extent that of acetylcholine 190. By contrast, other studies have shown that in the forearm of healthy volunteers and patients with various pathologies, the cytochrome P450 inhibitor sulfaphenazole does not influence basal tone or does not reduce bradykinin‐induced vasodilatations when given alone or in the presence of inhibitors of cyclooxygenase and NO synthase 191–193. Furthermore, in patients with coronary artery disease, inhibition of cytochrome P450 2C9 paradoxically improves the endothelium‐dependent, nitric oxide‐mediated vasodilatation 193, most likely because cytochrome P450 2C9 generates superoxide anions which reduce the bioavailability of NO 194. The discrepancies between these various studies can possibly be explained by two different phenomena. Residual NO is an important issue in human studies performed in vivo where typically only relatively low doses of one of the weakest inhibitors of NO synthase (L‐NMMA) can be administered. Furthermore, the lack of selectivity of the available inhibitors of cytochrome P450 makes it difficult to determine whether or not the release of EETs underlies non‐NO‐, non‐PGI₂‐mediated responses. For instance, inhibitors of cytochrome P450 of the imidazole family block various populations of K⁺ channels and especially IK_Ca, the K⁺ channel involved in EDHF‐mediated responses 195.

Therefore, in humans, responses resistant to the inhibition of NO synthases and cyclooxygenases can occur in vivo, and they contribute to the regulation of basal blood flow as well as to the endothelium‐dependent dilatations in response to flow and certain neurohumoral mediators.

Endothelium‐dependent responses resistant to inhibitors of NO synthases and cyclooxygenases: endothelial dysfunction and therapeutic interventions

Endothelial dysfunction, observed in various cardiovascular diseases, is associated with a decrease of NO synthesis and/or a loss of its biological activities. However, alteration of the EDHF pathway can also contribute to these endothelial dysfunctions or conversely can compensate the loss of NO bioavailability 1,2, 196.

The alteration of EDHF‐mediated responses has been reported with aging and various pathological conditions (hypertension, atherosclerosis, hypercholesterolemia, heart failure, ischemia‐reperfusion, angioplasty, eclampsia, diabetes, sepsis) 1,196. However, again, depending on the model or the vascular bed studied, marked differences can be observed. Hypertension per se does not produce a consistent depression of the EDHF‐mediated responses. For instance, in the mesenteric artery of the spontaneously hypertensive rat (SHR), the impairment of the endothelium‐dependent relaxations is attributed to a marked attenuation of the EDHF component and a concomitant production of cyclooxygenase‐derived contractile prostanoids (endothelium‐derived contractile factors, EDCF) with no or little alteration in the production of NO, whereas, in the model of L‐arginine‐methyl‐ester (L‐NAME)‐induced hypertension and in the same vascular bed, an increase in EDHF‐mediated responses may compensate the inhibition of NO production 1. In human with essential hypertension, the mechanism underlying the endothelial dysfunction is also linked to oxidative stress and the activation of cyclooxygenase, which reduces availability of NO. In the forearm vascular bed, the presence of an EDHF‐like pathway, possibly a cytochrome P450‐dependent mechanism, compensates the decreased NO bioavailability in order to sustain endothelium‐dependent vasodilatation 190,197. This compensatory pathway can be depressed by additional aggravating factors such as hyperhomocysteinemia 197. In myometrial arterioles from preeclamptic mothers, the upregulation of the EDHF‐mediated responses observed in normal pregnancy does not occur 198. The question remains whether the impairment of EDHF‐mediated responses contributes to the genesis of the syndrome or is a consequence of the hypertensive process.

Hypercholesterolemia is generally associated with a preserved or an enhanced contribution of EDHF‐mediated responses that compensate for the decrease in NO‐mediated relaxation 199. The resistance of endothelium‐dependent hyperpolarizations to hypercholesterolemia has been demonstrated in arteries from rabbit 200, SHR 201, apolipoprotein E‐deficient (ApoE‐/‐) mice 202–204 and dyslipidemic transgenic mice expressing human apolipoprotein B (hApoB+/+) 205. However, in isolated gastroepiploic arteries from atherosclerotic patients, endothelium‐dependent hyperpolarizations are inhibited 172. The prolonged duration of hypercholesterolemia and the severity of the atherosclerotic process in the human may contribute to the degree of dysfunction of the EDHF pathway 1.

Conversely, EDHF‐mediated responses are depressed in various models of type I and type II diabetes, with the exception of some murine models 196. In patients with type I diabetes under good glycemic control and without albuminuria, endothelial function appears normal, and both the NO‐ and the EDHF‐mediated responses are preserved. However, in patients with microalbuminuria, impairment of the endothelium‐dependent vasodilatation is observed 206. In these patients and in patients with type II diabetes, whether or not the various components of the endothelium‐dependent vasodilatation are differentially affected by the disease is not yet known 1,206.

No drug is available which has been designed to target EDHF‐mediated responses. Nevertheless, therapeutic interventions, with beneficial effects on the cardiovascular system such as angiotensin‐converting enzyme inhibitors, antagonists of angiotensin receptors, and phosphodiesterase‐3 inhibitors 1,207 can restore these responses, suggesting that the improvement of the EDHF pathway contributes to the observed beneficial effect. Similarly, various so‐called nonpharmacological therapeutic strategies, including exercise and supplementation with estrogens, omega‐3 polyunsaturated fatty acids, polyphenol derivatives, potassium and/or calcium, help to reverse endothelial dysfunction including blunted EDHF‐mediated responses. Whether or not the improvement of these EDHF‐mediated responses contributes to the beneficial effects of these dietary maneuvers can be suspected but is far from being demonstrated 1, 198.

Conclusions and perspectives

Endothelial cells control the tone of the underlying vascular smooth muscle by releasing numerous vasoactive substances, including NO, reactive oxygen species, potassium ions, and metabolites of arachidonic acid (e.g. prostacyclin, EETs, lipoxygenase derivatives). Furthermore, the endothelial monolayer behaves as a conductive tissue propagating an electrical signal along the axis of the blood vessel by means of homocellular gap junctions and throughout the vascular wall itself by means of myoendothelial gap junctions. Endothelium‐dependent relaxations, independent of the production of NO and prostacyclin, probably play an important role in cardiovascular physiology in numerous animal species and in the human. They can act as a backup system when NO is inhibited or reduced, but this is not necessarily the case.

However, it is often difficult to reach a conclusion as to the true importance of endothelium‐dependent hyperpolarizations because of the use of unspecific pharmacological tools and the lack of electrophysiological measurements. The mechanism underlying endothelium‐responses must be carefully dissected in order to be properly identified. The synthesis of more selective compounds such as nonpeptidic inhibitors and activators of SK_Ca and IK_Ca208,209, may in the future allow the selective blockade/activation of EDHF‐mediated responses and hence the proper determination of their physiological role in the human circulation.

The improvement or restoration of EDHF responses has not been, yet, the direct purpose of any pharmaceutical effort. Activating or increasing the expression of endothelial TRP, IK_Ca and/or SK_Ca channels, and smooth muscle K_IR and/or specific isoform(s) of Na⁺/K⁺‐ATPase, as well as facilitating myoendothelial communication (rotigaptide‐like drugs, phosphodiesterase‐3 inhibitors) and increasing the expression of appropriate connexins (Cx40, Cx43, Cx37) represent new potential targets.